The NOAA Operational Numerical Guidance System: EMC overview

Transcription

1 The NOAA Operational Numerical Guidance System: EMC overview Hendrik Tolman, Michael Ek et al Environmental Modeling Center (EMC) NOAA/NWS/NCEP CICS/CPC and EMC mini-workshop at ESSIC College Park, Maryland, USA, 6 May

2 Outline NCEP and EMC Production Suite Components: o Data Assimilation o Global Forecast System (GFS), Global Ensemble (GEFS) o Unified Global Coupled System (medium-range to seasonal) o North American Multi Model Ensemble (Seasonal) o North America Mesoscale (NAM) model, Hurricane (HWRF), Short-range Ensemble Forecast (SREF) o Hurricane, Land-Hydrology, Ocean, Sea-ice o NOAA Environmental Modeling System (NEMS) NCEP/EMC Model Evaluation Group (MEG) 2

3 National Centers for Environmental Prediction Specialized Services Common Mission 6 NOA Corps Officers NHC: Brinkley EMC: Ostapenko OPC: Schultz (Phillips) AWC: Waddington SWPC: Hemmick OD: Odell Environmental Modeling Center College Park, MD Climate Prediction Center College Park, MD Aviation Weather Center Kansas City, MO National Hurricane Center Miami, FL NCEP Central Operations College Park, MD (Supercomputers in Reston & Orlando) Ocean Prediction Center College Park, MD Space Weather Prediction Center Boulder, CO Storm Prediction Center Norman, OK Mission: Reliable, timely, and accurate analyses, guidance, forecasts, and warnings for the protection of life and property and the enhancement of the national economy. Vision: Nation s trusted source, first alert, and preferred partner for environmental prediction services. EMC: Enhancement, transition, and maintenance of numerical models that provide operational guidance. Weather Prediction Center College Park, MD 3

4 NOAA s Operational Numerical Guidance Suite (Feb 2015) 3D-VAR DA 3D-En-Var DA Climate Forecast System (CFS) GFS, MOM4, GLDAS/LIS/Noah, Sea Ice Global Forecast System (GFS) Global Spectral Noah LSM Global Ensemble Forecast System (GEFS) 21 GFS Members North American Ensemble Forecast System GEFS, Canadian Global Model 3D-VAR DA Regional Hurricane GFDL WRF-NMM 3D-VAR DA Regional NAM NMMB Noah land model Short-Range Ensemble Forecast 21members WRF (ARW + NMM) NMMB 7members each High Res Windows WRF-ARW & NMMB Waves WaveWatch III Ocean (RTOFS) HYCOM SURGE SLOSH P-SURGE SLOSH ESTOFS ADCIRC 3D-VAR DA Dispersion HYSPLIT Air Quality CMAQ Ecosystem EwE Regional Bays Great Lakes (POM) N Gulf of Mexico (FVCOM) Columbia R. (SELFE) Chesapeake (ROMS) Tampa (ROMS) Delaware (ROMS) Rapid Refresh WRF ARW North American Land Surface Data Assimilation System Noah Land Surface Model NEMS Aerosol Global Component (NGAC) GFS & GOCART Space Weather ENLIL 3D-VAR DA High-Res RR (HRRR) WRF ARW 4

5 Data Assimil. Plans: Next Global Implementation (~Dec. 2015) Hybrid 4D EnVar: Hourly time bins (7 time levels per analysis). 80 ensemble members. T1534L64 deterministic with T574L64 ensembles used for T574 analysis increments Modified thinning of data in time. Observation changes: Cloudy Microwave Radiances: Cloud analysis impacted by covariances from Hybrid and cloud impacted microwave radiances. Upgraded CRTM2.2. Additional aircraft and AMV data. Bias correction of aircraft temperature observations. NSST: Variation in Near Sea-Surface Temperature of ocean included in both forecast model and analysis. Directly uses radiances and accounts for depth of observations in NSST analysis. Diurnal cycle of Ocean and observations accounted for much better. John Derber 5

6 Global Model Plans Q2FY15 implementation (January 14, 2015) T1534 Semi-Lagrangian (~13km) Use of high resolution daily SST and sea ice analysis High resolution until 10 days Physics: - Radiation modifications - Reduced drag coefficient at high wind speeds - Stationary convective gravity wave drag - Soil Moisture climatology from CFSv2 - Changes to roughness length calculations FY16 NEMS 4D-Hybrid EnVAR Convective upgrade Land and surface layer upgrade FY17 T1534L64 > T1534L128 Enhanced physics Aerosol prediction initially lower resolution used as forcing in high res. WAM Whole Atmosphere Model initially lower resolution up to 600km Mark Iredell 6

7 Global Ensemble Plans Current Next GEFS Configuration (v11.0.0) Model GFS Euler model (V9.0.1) GFS Semi-Lagrangian model (V10.0.0) Horizontal res. T254 (~52km) for hours, T190 (~70km) for hrs TL574 (~34km) for hours, TL382(~52km) for hr Vertical res. L42 hybrid levels L64 hybrid levels to match GFS and DA Compute 84 nodes (+ post process) for 55 minutes 300 nodes (first 35 minutes), 250 nodes (2nd 30 minutes) Output every 6-hr for 1*1 degree pgrb files every 3-hr for 0.5*0.5 degree pgrb files Schedule Feb deliver codes/scripts to NCO Apr implementation (WCOSS-phase II) Increase model resolution and membership. Introduce other stochastic schemes: - Stochastic Kinetic Energy Backscatter (SKEB) represents processes absent from model. - Stochastic Perturbed Physics Tendencies (SPPT) designed to represent the structural uncertainty (or random errors) of parameterized physics. Biggest impact for tropics. - Stochastically-perturbed boundary layer HUMidity (SHUM) designed to represent influence of sub-grid scale humidity variability on the the triggering of convection. - Stochastic perturbed land-surface. Extend GEFS to 35 days: Coupling with ocean/ice model, or alternate method. Yuejian Zhu 7

8 NCEP Unified Global Coupled Hybrid Data Assimil. and Forecasting Suite: Atmosphere, Land, Ocean, Sea-ice, Wave, NEMS Chemistry, Ionosphere NEMS OCEAN SEA-ICE WAVE LAND AERO ATMOS ATMOS AERO LAND WAVE SEA-ICE OCEAN Coupled Model Ensemble Forecast Coupled Model Ensemble Forecast Coupled Ensemble Forecast (N members) Ensemble Analysis (N Members) INPUT OUTPUT Suru Saha 8

9 NMME = National (or North American) Multi- Model Ensemble for Seasonal Prediction Real-Time definition: Each month a global NMME seasonal forecast of SST, precip and T2m on a global 1X1 grid in support of CPC operations. (This is phase I ) In early 2014, the SIX real time participants were CFSv2, NCAR- UoFlorida-CCSM3, NASA-GEOSS, GFDL-CM2, Canadian CMC3 and CMC4. (Retired participants include CFSv1, IRI-f and IRI-a) During 2014, two models were added: NCAR-UoFlorida-CCSM4, and GFDL-Flor-a/b. A third model NCAR-CESM1 is in the pipeline (hindcasts 1982-present to be finished). CCSM3 has been retired by late Currently SEVEN models, soon EIGHT. Close to 100 ensemble members in hindcasts. More members in real time. A review of NMME took place Sept 2014 and NMME as a CPUdistributed system will be continued for seasonal prediction and is considered operational. The phase II R&D effort (daily data to study intra-seasonal) has yielded a big archive managed at NCAR. The IMME carries on each month. IMME at NCEP ~ Eurosip (at ECMWF). Huug van den Dool, CPC 9

10 Recent and planned upgrades to North American Mesoscale (NAM) system NAM 2014Q4 upgrade NAM forecast configuration: - 12km parent to 84-hr. - 4 fixed nests: 4km CONUS, 6km Alaska, 3km HI/PR to 60-hr km relocatable storm scale nest to 36-hr. Hybrid variational ensemble analysis; new satellite bias correction algorthim. RRTM radiation scheme. Ferrier-Aligo microphysics; tuned to improve severe storm structure. Convection: Modified BMJ (moister profiles) to improve 12km parent bias; explicit in nests except AK. NAM 2016Q1 upgrade CONUS/Alaska nest to 3km, explict convection in AK nest. New shallow convection scheme in 12km parent NAM; improves cold season QPF bias. More frequent calls to radiation/physics. Digital filter with radar-derived temperature tendencies. 4-d EnKF. Rapidly updated hourly assimilation system (NAMRR). Others T.B.D. Eric Rogers 10

11 June 29-30, 2012 DC Derecho - 27hr fcst NAMRR Test with 3km CONUSnest Significant improvement at longer lead times with 3km NAMRR relative to Operations (at the time) 4 km Ops NAM CONUSnest Fhr=27 3 km NAMRR CONUSnest Fhr=27 Observations Eric Rogers 11

12 Moore OK Severe Weather Event 20 May 2013 Ops 4km NAMRR 3km Obs Fhr = 03 Valid 21Z Fhr = 16 Valid 22Z Eric Rogers 12

13 Rapid Refresh and HRRR NOAA hourly updated models (situational awareness for energy, aviation, severe weather, etc.) 13km Rapid Refresh (RAP) (mesoscale) Version 2 -- NCEP implement 25 Feb 2014 Version 3 GSD Planned NCEP summer km High Resolution Rapid Refresh (HRRR) (storm-scale) Initial -- NCEP implement 30 Sept 2014 Version 2 GSD Planned NCEP summer 2015 RAP HRRR 13

14 NCEP RAPv3 and HRRRv2 Model Run at: Domain Grid Points Grid Spacing Vertical Levels Pressure Top Boundary Conditions Initialized RAP GSD, NCO North America 758 x km mb GFS Hourly (cycled) HRRR GSD CONUS 1799 x km mb RAP Hourly - RAP (no-cycle) Model Version Assimilation Radar DA Radiation LW/SW Microphysics Cumulus Param PBL LSM RAP WRF-ARW v3.6+ GSI Hybrid 3D- VAR/Ensemble 13-km DFI RRTMG/RR TMG Thompson aerosol v3.6.1 GF v3.6+ MYNN v3.6+ RUC v3.6+ HRRR WRF-ARW v3.6+ GSI Hybrid 3D- VAR/Ensemble 3-km 15-min LH RRTMG/ RRTMG Thompson aerosol v3.6.1 GF shallow MYNN v3.6+ RUC v3.6+ Model Horiz/Vert Advection Scalar Advection Upper-Level Damping 6 th Order Diffusion SW Radiation Update Land Use MP Tend Limit Time- Step RAP 5 th /5 th Positive- Definite w-rayleigh 0.2 Yes min MODIS Fractional 0.01 K/s 60 s HRRR 5 th /5 th Positive- Definite w-rayleigh 0.2 Yes 0.25 (flat terr) 15 min with SWdt (Ruiz-Arias) MODIS Fractional 0.07 K/s 20 s 14

15 HRRRv2 12z 25 Mar h fcst Valid 23z 15

16 Current ~2015 ~2018 SREF continental scale SREF continental scale SREF continental scale WRF-ARW, -NMM, NMMB WRF-ARW & NMMB WRF-ARW & NMMB 7 each = 21 members 16km 35 levels 6 hourly to 87 hr NCEP Mesoscale Ensembles Replace Regional Deterministic Guidance: Current and Future Convection-Allowing- Scale hrly 3km HRRR & NAM nest run to 15 h for CONUS Irregular suite of guidance 3-6km, ~6 hourly to 48/60 hr for CONUS, Alaska, HI, PR 13 each = 26 members ~16km 40 levels 6 hourly to 87 hr NARRE-TL run hourly to 18 hr Convection-Allowing-Scale Upgrade irregular suite (HiResW) to ~3km, 6 hourly to 48/60 hr for CONUS, Alaska, HI, PR 13 each = 26 members ~9km (parent) levels 6 hourly to 96 hr. SREF_RR run hourly to 24 hr Convection-Allowing- Scale Ensemble HREF_RR (high-res ensemble fcst): Multiple hourly 3km run to 24 hr; HREF: run 6 hourly extended to 60 hr for CONUS, Alaska, Hi, PR Storm Scale Storm Scale Storm Scale Ensemble Single placeable sub-nest [fire weather run] km Run 6 hourly to 36 hr Single placeable/movable subnest 1-1.5km Run 6 hourly to 36 hr Storm-scale ensemble (SSE) Multiple placeable/movable sub-nests: ~1km run hourly to 18 hr and run 6 hourly to 36 hr Jun Du 16

17 NCEP Operational HWRF Model: NOAA s State-of-the-art Hurricane Prediction System HWRF Model Configuration: Three telescopic domains with high resolution storm following nests: 27km/9km/3km; 61L going to 18km/6km/2km in Coupled to MPI POM Ocean Model worldwide using NCEP Coupler. Sophisticated vortex initialization. EnKF/GSI hybrid regional DA system with real-time inner core aircraft data assimilation. Advanced physics tailored for hurricane conditions based on observations. Coupled Noah LSM in 2015 Expanding the scope of operational HWRF for all global tropical cyclones in Accurate and reliable Track, Intensity & Structure Forecasts Real-time HWRF products, all global TCs: Significant Improvements in the past 5 years, meeting HFIP Goals Vijay Tallapragada 17

18 Land Data Assimilation System (LDAS) 5 Aug 2014: North American LDAS (NLDAS) operational. NLDAS: 4 land models run uncoupled, driven by CPC observed precipitation & NCEP R-CDAS atmospheric forcing. Output: 1/8-deg. land & soil states, surface fluxes, runoff & streamflow; anomalies from 30-yr climatology for drought. Future: higher res. (~3-4km), extend to N.A./global domains, improved land data sets/data assimil. (soil moisture, snow), land model physics upgrades inc. hydro., initial land states for weather & seasonal climate models; global drought information. July 30-year climo. July 2011 September 2013 Ensemble monthly soil moisture anomaly Daily streamflow Mike Ek 18

19 Global and Basin Scale Ocean Forecast Systems Eddy Resolving Ocean Modeling and Initialization Coupled Modeling for Hurricanes (Air-Sea-Wave flux interaction, mixing) Coupled Modeling for short-term, medium-term and seasonal scales (Air-Sea-Ice-Wave flux interactions, ensembles) Coupled Ecosystem Forecasting (Biogeochemical, NPZD, tracers) All Ocean Forecast Systems presently based on HYCOM, RTOFS (Real Time Ocean Forecast Systems) represents line of products Primary Users: NOAA: NWS, NOS, IOOS, OAR, JCSDA External: DHS, Academia, Japan (JAEA), India (INCOIS) Strong collaboration with US Navy, leveraging core HYCOM and data assimilation developments at NRL. Avichal Mehra 19

20 WAVEWATCH III and Wave Coupling GFS model air-sea fluxes depend on sea state (roughness -> Charnock). WAVEWATCH III model forced by wind from GFS and currents from Ocean. Ocean model forced by heat flux, sea state dependent wind stress modified by growing or decaying wave fields and Coriolis-Stokes effect. Current Status: o Wave model driven in uncoupled mode with GFS 10m winds for global forecasts (4 cycles a day out to 180 h of forecast). o Hurricane wave model uses blend of GFS and HWRF winds for forecasts (4 cycles a day out to 126 h of forecast). o Use first guess model results from previous cycle as final wave analysis. Planned Upgrades: o Drive the wave model in a coupled mode with atmospheric winds (for wave dependent boundary conditions in atmospheric models). o Include wave-ocean coupling (currents from ocean model to wave model, and wave induced langmuir mixing & stokes drift from wave model to ocean). o Data assimilation of significant wave heights to develop a wave analysis: GSI and LETKF approaches. o Planned sources of data: Spectral data from ocean buoys; Satellite data from altimeters. Arun Chawla, Henrique Alves, Hendrik Tolman 20

21 NOAA Environmental Modeling System (NEMS) Unify NCEP operational systems under a single framework. More easily share common structures/components. Expedite interoperability. First two systems under NEMS implemented in NCEP ops: NAM, and NEMS Global Aerosol Component (NGAC). (MOM, HYCOM) Ionosphere Mark Iredell 21

22 Land-Atmosphere Interactions 22

23 Model Evaluation Group (MEG) Verifies and evaluates daily performance of EMC forecast/analysis systems from a synoptic and mesoscale perspective to complement the statistical perspective. Conducts weekly EMC map discussion of model performance - Audience recently expanded to involve more NWS regional and local forecasters. Project benefits: - Provides critical feedback to modelers and branch chiefs. - Provides streamlined feedback to outside users with model concerns. - In first 3 years, several model problems noted, brought to attention of EMC modeling team, and corrected or at least mitigated in operations. MEG also to work closely with forecasters to understand and evaluate proposed model upgrades prior to the short formal evaluation period. How can MEG best include experience and knowledge of model users? Glenn White, Geoff Manikin, Corey Guastini 23

24 Thank you! 24